From: | Peter Eisentraut <peter(at)eisentraut(dot)org> |
---|---|
To: | Tomas Vondra <tomas(dot)vondra(at)enterprisedb(dot)com>, Bruce Momjian <bruce(at)momjian(dot)us> |
Cc: | Andres Freund <andres(at)anarazel(dot)de>, David Christensen <david(dot)christensen(at)crunchydata(dot)com>, PostgreSQL-development <pgsql-hackers(at)postgresql(dot)org>, Stephen Frost <sfrost(at)snowman(dot)net> |
Subject: | Re: Initdb-time block size specification |
Date: | 2023-07-04 15:17:55 |
Message-ID: | 747c434a-0931-6f41-dd92-281f2d8f07c2@eisentraut.org |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-hackers |
On 01.07.23 00:21, Tomas Vondra wrote:
> Right, that's the dance we do to protect against torn pages. But Andres
> suggested that if you have modern storage and configure it correctly,
> writing with 4kB pages would be atomic. So we wouldn't need to do this
> FPI stuff, eliminating pretty significant source of write amplification.
This work in progress for the Linux kernel was also mentioned at PGCon:
<https://lwn.net/Articles/933015/>. Subject the various conditions, the
kernel would then guarantee atomic writes for blocks larger than the
hardware's native size.
From | Date | Subject | |
---|---|---|---|
Next Message | Heikki Linnakangas | 2023-07-04 15:33:05 | Re: Creation of an empty table is not fsync'd at checkpoint |
Previous Message | Dave Cramer | 2023-07-04 15:05:45 | Re: Why is DATESTYLE, ordering ignored for output but used for input ? |